NLU only and different configurations

Hello, I am trying to use NLU only mode, and from documentation here it seems only the default configuration will be used. Unfortunately, the results I am getting trying to extract simpler regex id entity, are far from perfect, and seems like the DIETClassifier is the only extractor that get into account, while I do define the entity regex. I already verified my regex are correct.

Is there a better way to configure this ? Is there a way to “force” different configuration for NLU only mode ?

Please advise, Shushu nlu.yml (5.1 KB)

You’ll need to uncomment the config & add the regexentityextractor to your NLU pipeline: Components

Thanks. The documentation I read was misleading, the the output showing the entity is coming from the DIETClassifier was not what I expected. In the bottom line, once I played with the config.yml, I saw it effect the train nlu action.